Kevin Hayes: Knowledge from the Glass Age Page 2
We make a derivative of the PA80/80, called the PA35/35 ($2950), which changes the output tubes to EL34 pentodes and takes both the suppressor grid and the screen grid directly to the anode, forming a real triode structure—which is different from the way almost everyone (including us, in our earlier designs) does it. That amplifier has variable grounding and variable feedback, but in that circuit, the variable feedback is far more useful than it would be on the PA80/80.
Phillips: What was your design brief for the PA80/80?
Hayes: The PA80/80 stems from a product we made called the Vintage Williamson, which was a 35Wpc power amp. It was a nice-sounding amp, but it missed in the American market for a couple of reasons. First, its power rating was below what many in this market feel they require—remember, this was before the big SE revival of 10Wpc amplifiers—and second, we designed it with a full-height faceplate that obscured the tubes from view. For those reasons, I feel, the product fell by the wayside, but it had some qualities that I felt were just too nice to ignore. The PA80/80 retains those qualities, but offers greater power, control, and dynamics. It also offers better cosmetics and user convenience.
I guess the design brief on the 80/80 really comes down to this: seven years ago, I couldn't get happy with anything at the $3000 price point. I wanted to create something that made a contribution in that range—something that I, if I were in that position today, would gobble up in a heartbeat. And that was it, really. I wanted to make something that would satisfy me—at a price I could afford if I didn't own VAC.
As I talked with Kevin Hayes, we touched on many subjects, including his engineering background—or lack thereof—when he started the Valve Amplification Company. Rather than being defensive about it, he argued that it may have, in fact, been advantageous.
Hayes: When I started the company, my electronics background was almost entirely informal. My bachelor's was in experimental psychology—the perceptual-encoding side of psychology—and my master's was in business, manufacturing actually. My doctorate was in a field of mathematics called operations research—you'd be amazed at how much of all this actually has proven useful. But in terms of actual circuit information, I was primarily self-taught.
That's not a bad thing, because the traditional electronics curriculum of the last 20 years does not encompass any information about the vacuum tube. During the years I was at Duke, I would go into the Engineering Library archives and spend time looking through the old publications, such as the RCA Review going back into the 1920s—I believe that we've forgotten more about tube technology than we currently remember. I found some fascinating things in those journals that are no longer common knowledge.
Phillips: Such as the "sleeping sickness," which develops in tubes kept in soft-start mode?
Hayes: That's the informal name for it, yes. That piece of data was developed during the 1950s, and it came out of two non-audio-related fields. The first was radar, where you put a very brief, very powerful pulse through the circuit after a long period of rest, and the other was computers, where the tube acted as an on/off switch. The tube may be turned off, with no plate current flowing, for a long time and then suddenly be asked to do something. If you do have the tube hot, but not flowing plate current, a change happens to the cathode—it builds up an interference layer, which acts in practice like a capacitor in series with the cathode, and you can't do anything about it. You'd be unlikely to find that information, if you only searched audio-only references, however.
Phillips: Does the interference layer become a permanent function of the tube, then?
Hayes: That's an unknown. The tube does seem to rejuvenate over time, but whether it's complete or not remains undetermined. Given how dynamic the cathode is, however, I'd be inclined to mess with it as little as possible.
When you go back and examine the physical and chemical sciences of tube construction, there are many ways that people have gone about building them: different cathode lead materials, cathode coatings, getter materials, methods of evacuation—all of these react differently over time. One particular brand of 12AX7 may be more prone to cathode interference than another, but we aren't certain, since serious tube research more or less ended with the introduction of the transistor—before we'd fully answered all the questions.